rtestim: Time-varying reproduction number estimation with trend filtering DOI Creative Commons

Jiaping Liu,

Zhenglun Cai, Paul Gustafson

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Dec. 18, 2023

To understand the transmissibility and spread of infectious diseases, epidemiologists turn to estimates instantaneous reproduction number. While many estimation approaches exist, their utility may be limited. Challenges surveillance data collection, model assumptions that are unverifiable with alone, computationally inefficient frameworks critical limitations for existing approaches. We propose a discrete spline-based approach solves convex optimization problem---Poisson trend filtering---using proximal Newton method. It produces locally adaptive estimator number heterogeneous smoothness. Our methodology remains accurate even under some process misspecifications is efficient, large-scale data. The implementation easily accessible in lightweight R package rtestim (dajmcdon.github.io/rtestim/).

Language: Английский

rtestim: Time-varying reproduction number estimation with trend filtering DOI Creative Commons

Jiaping Liu,

Zhenglun Cai, Paul Gustafson

et al.

PLoS Computational Biology, Journal Year: 2024, Volume and Issue: 20(8), P. e1012324 - e1012324

Published: Aug. 6, 2024

To understand the transmissibility and spread of infectious diseases, epidemiologists turn to estimates instantaneous reproduction number. While many estimation approaches exist, their utility may be limited. Challenges surveillance data collection, model assumptions that are unverifiable with alone, computationally inefficient frameworks critical limitations for existing approaches. We propose a discrete spline-based approach solves convex optimization problem—Poisson trend filtering—using proximal Newton method. It produces locally adaptive estimator number heterogeneous smoothness. Our methodology remains accurate even under some process misspecifications is efficient, large-scale data. The implementation easily accessible in lightweight R package rtestim .

Language: Английский

Citations

0
Interpreting epidemiological surveillance data: A modelling study from Pune City DOI Creative Commons

Prathith Bhargav,

Soumil Kelkar, Joy Merwin Monteiro

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 14, 2024

Abstract Routine epidemiological surveillance data represents one of the most continuous and comprehensive sources during course an epidemic. This is used as inputs to forecasting models well for public health decision making such imposition lifting lockdowns quarantine measures. However, generated testing contact tracing not through randomized sampling which makes it unclear how representative epidemic itself. Using BharatSim simulation framework, we build agent-based model with a detailed algorithm actual strategies employed in Pune city generate synthetic data. We simulate impact different strategies, availability tests efficiencies on resulting The fidelity representing real-time state decision-making explored context city.

Language: Английский

Citations

0
A Flexible Framework for Local-Level Estimation of the Effective Reproductive Number in Geographic Regions with Sparse Data DOI Creative Commons

Md Sakhawat Hossain,

RK Goyal, Natasha K. Martin

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 7, 2024

Abstract Background Our research focuses on local-level estimation of the effective reproductive number, which describes transmissibility an infectious disease and represents average number individuals one person infects at a given time. The ability to accurately estimate in geographically granular regions is critical for disaster planning resource allocation. However, not all have sufficient outcome data; this lack data presents significant challenge accurate estimation. Methods To overcome challenge, we propose two-step approach that incorporates existing R t procedures (EpiEstim, EpiFilter, EpiNow2) using from geographic with (step 1), into covariate-adjusted Bayesian Integrated Nested Laplace Approximation (INLA) spatial model predict sparse or missing 2). flexible framework effectively allows us implement any procedure coarse entirely data. We perform external validation simulation study evaluate proposed method assess its predictive performance. Results applied our South Carolina (SC) counties ZIP codes during first COVID-19 wave (‘Wave 1’, June 16, 2020 – August 31, 2020) second 2’, December March 02, 2021). Among three methods used step, EpiNow2 yielded highest accuracy prediction Median county-level percentage agreement (PA) was 90.9% (Interquartile Range, IQR: 89.9-92.0%) 92.5% (IQR: 91.6-93.4%) Wave 1 2, respectively. zip code-level PA 95.2% 94.4-95.7%) 96.5% 95.8-97.1%) Using EpiEstim, ensemble-based median ranging 81.9%-90.0%, 87.2%-92.1%, 88.4%-90.9%, respectively, across both waves granularities. Conclusion These findings demonstrate methodology useful tool small-area , as yields high

Language: Английский

Citations

0
rtestim: Time-varying reproduction number estimation with trend filtering DOI Creative Commons

Jiaping Liu,

Zhenglun Cai, Paul Gustafson

et al.

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Dec. 18, 2023

To understand the transmissibility and spread of infectious diseases, epidemiologists turn to estimates instantaneous reproduction number. While many estimation approaches exist, their utility may be limited. Challenges surveillance data collection, model assumptions that are unverifiable with alone, computationally inefficient frameworks critical limitations for existing approaches. We propose a discrete spline-based approach solves convex optimization problem---Poisson trend filtering---using proximal Newton method. It produces locally adaptive estimator number heterogeneous smoothness. Our methodology remains accurate even under some process misspecifications is efficient, large-scale data. The implementation easily accessible in lightweight R package rtestim (dajmcdon.github.io/rtestim/).

Language: Английский

Citations

0